RLTasksby ulam.ai
Models / OpenAI

OpenAI / MODEL EVIDENCE

GPT-5.2

77 benchmarks · 140 observations · 67 sources · released 2025-12-11

Scores at a glance

77 benchmarks · large dot: GPT-5.2; small dots: closest models where reported

MCPMark ↗57.5
29.557.5

GPT5.2 · MCP-Mark · original release · Qwen report ↗

Terminal-Bench 2.0 ↗54.0
22.559.3

GPT5.2 · Terminal Bench 2 · Qwen report ↗

6 more reported results

64.94% · GPT-5.2 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

62.92% · GPT-5.2 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

60.67% · GPT-5.2 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

54.04% · GPT-5.2 (unknown thinking) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

64.9% · GPT-5.2 (medium) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

62.9% · GPT-5.2 (medium) · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗

GPQA Diamond ↗91.67 %
27.0295.45

GPT 5.2 · GPQA Diamond · Accuracy ↗

4 more reported results

91.4% · GPT-5.2 (xhigh) · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗

88.19% · GPT-5.2 (high) · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗

87.88% · GPT-5.2 (medium) · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗

82.7% · GPT-5.2 (low) · GPQA diamond · Best score (across scorers) · % · Epoch AI · GPQA diamond ↗

FrontierMath Tiers 1-3 (v1, old) ↗40.7%
40.740.7

GPT-5.2 (xhigh) · FrontierMath-2025-02-28-Private · Best score (across scorers) · % ↗

3 more reported results

40.34% · GPT-5.2 (high) · FrontierMath-2025-02-28-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-2025-02-28-Private ↗

36.9% · GPT-5.2 (medium) · FrontierMath-2025-02-28-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-2025-02-28-Private ↗

26.55% · GPT-5.2 (low) · FrontierMath-2025-02-28-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-2025-02-28-Private ↗

FrontierMath Tier 4 (v1, old) ↗18.8%
18.818.8

GPT-5.2 (xhigh) · FrontierMath-Tier-4-2025-07-01-Private · Best score (across scorers) · % ↗

3 more reported results

18.8% · GPT-5.2 (high) · FrontierMath-Tier-4-2025-07-01-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-Tier-4-2025-07-01-Private ↗

16.7% · GPT-5.2 (medium) · FrontierMath-Tier-4-2025-07-01-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-Tier-4-2025-07-01-Private ↗

6.25% · GPT-5.2 (low) · FrontierMath-Tier-4-2025-07-01-Private · Best score (across scorers) · % · Epoch AI · FrontierMath-Tier-4-2025-07-01-Private ↗

OTIS Mock AIME 2024-2025 ↗96.11%
96.1196.11

GPT-5.2 (high) · OTIS Mock AIME 2024-2025 · Best score (across scorers) · % ↗

3 more reported results

96.1% · GPT-5.2 (xhigh) · OTIS Mock AIME 2024-2025 · Best score (across scorers) · % · Epoch AI · OTIS Mock AIME 2024-2025 ↗

93.89% · GPT-5.2 (medium) · OTIS Mock AIME 2024-2025 · Best score (across scorers) · % · Epoch AI · OTIS Mock AIME 2024-2025 ↗

78.89% · GPT-5.2 (low) · OTIS Mock AIME 2024-2025 · Best score (across scorers) · % · Epoch AI · OTIS Mock AIME 2024-2025 ↗

GSO ↗27.45%
9.888.24

GPT-5.2 (high) · GSO-Bench · Score OPT@1 · % ↗

LMCA ↗43.91%
2.7868.26

GPT-5.2 (xhigh) · LMCA · Score · % ↗

DTBench ↗90.93%
41.6598.93

GPT-5.2 (xhigh) · DTBench · Accuracy · % ↗

EnigmaEval ↗10.39%
0.3839.28

GPT-5.2 (unknown thinking) · EnigmaEval · Accuracy · % ↗

Vals RSI Index ↗0.29 reference score
00.46

GPT 5.2 (xhigh) · Compression · reference-anchored score ↗

3 more reported results

0.31 reference score · GPT 5.2 (xhigh) · LM training · reference-anchored score · Vals RSI Index · LM training ↗

0.04 reference score · GPT 5.2 (xhigh) · Harness engineering · reference-anchored score · Vals RSI Index · Harness engineering ↗

0.00 reference score · GPT 5.2 (xhigh) · Post-training · reference-anchored score · Vals RSI Index · Post-training ↗

OmniDocBench ↗86.59 points
78.4496.91

GPT-5.2 · OmniDocBench · Overall parsing score · end-to-end evaluation ↗

EconBench ↗10.4448 error points
10.4410.44

gpt-5.2 · EconBench · EconBench prototype · risk-consistency error · published prototype observations ↗

Closest models by score

Similar positions on shared benchmark leaderboards. Smaller gaps are closer.

16.4 percentile-point gap

21 of 27 shared benchmark comparisons

How similarity is measured

Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.

Filter comparisons to this model →

Reports featuring this model

Qwen3.5 · earlier-2026 baseline · Qwen →

Sources

Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.

  1. ARC Prize - Leaderboard ↗

    ARC Prize Foundation

  2. Arena leaderboard — WebDev evaluation ↗

    Arena

  3. AIME 2025 Benchmark | Artificial Analysis ↗

    Artificial Analysis

  4. Global-MMLU-Lite Benchmark | Artificial Analysis ↗

    Artificial Analysis

  5. LiveCodeBench Benchmark | Artificial Analysis ↗

    Artificial Analysis

  6. MMLU-Pro Benchmark | Artificial Analysis ↗

    Artificial Analysis

  7. Terminal-Bench Hard Benchmark | Artificial Analysis ↗

    Artificial Analysis

  8. TasteVal: Measuring the Experimental Research Taste of AI Systems Against Human Experts — published result table ↗

    Jaffe, Oliver, Sherburn, Dane · 2026

  9. DTBench — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  10. LMCA — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  11. EconBench: Economic choice by language models — gpt-5.2 prototype metrics ↗

    EconBench project · 2026

  12. Algotune — methodology and published results ↗

    Epoch AI · 2026

  13. ARC-AGI-1 ↗

    Epoch AI

  14. ARC-AGI-2 ↗

    Epoch AI

  15. Chess Puzzles ↗

    Epoch AI

  16. CL-bench ↗

    Epoch AI

  17. Earthborne Rangers (EBR-bench) ↗

    Epoch AI

  18. FrontierMath Tier 4 (v1, old) ↗

    Epoch AI

  19. FrontierMath Tier 4 (v2) ↗

    Epoch AI

  20. FrontierMath Tiers 1-3 (v1, old) ↗

    Epoch AI

  21. FrontierMath Tiers 1-3 (v2) ↗

    Epoch AI

  22. Furniture Assembly Benchmark ↗

    Epoch AI

  23. GPQA Diamond ↗

    Epoch AI

  24. Humanity's Last Exam ↗

    Epoch AI

  25. METR Time Horizons ↗

    Epoch AI

  26. Mystery Game Puzzles ↗

    Epoch AI

  27. OTIS Mock AIME 2024-2025 ↗

    Epoch AI

  28. Remote Labor Index ↗

    Epoch AI

  29. SimpleQA Verified ↗

    Epoch AI

  30. SimpleBench ↗

    Epoch AI

  31. SWE-bench Verified ↗

    Epoch AI

  32. Terminal-Bench 2.0 ↗

    Epoch AI

  33. Capabilities & benchmarking — retained external evaluation records ↗

    Epoch AI · 2026

  34. Vending Bench 2 — methodology and published results ↗

    Epoch AI · 2026

  35. Text Arena (Coding) ↗

    Epoch AI

  36. AI benchmarks & capabilities ↗

    Epoch AI

  37. FutureSearch Evals ↗

    FutureSearch

  38. GSO: Challenging Software Optimization Tasks for Evaluating SWE-Agents ↗

    Shetty, Manish, Jain, Naman, Liu, Jinjian, Kethanaboyina, Vijay, Sen, Koushik, Stoica, Ion · 2025

  39. WeirdML ↗

    Håvard Tveit Ihle

  40. Qwen/Qwen3.5-397B-A17B · Hugging Face ↗

    Qwen Team

  41. DrugDiscoveryBench — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  42. EnigmaEval ↗

    Scale Labs / SEAL

  43. Scale Labs — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  44. Metr Autonomy Task Horizon Suite — methodology and published results ↗

    METR · 2026

  45. OmniDocBench: Benchmarking Diverse PDF Document Parsing with Comprehensive Annotations — published result table ↗

    Linke Ouyang, Yuan Qu, Hongbin Zhou, Jiawei Zhu, Rui Zhang, Qunshu Lin et al. · 2024

    All 20 authors

    Linke Ouyang, Yuan Qu, Hongbin Zhou, Jiawei Zhu, Rui Zhang, Qunshu Lin, Bin Wang, Zhiyuan Zhao, Man Jiang, Xiaomeng Zhao, Jin Shi, Fan Wu, Pei Chu, Minghao Liu, Zhenxiang Li, Chao Xu, Bo Zhang, Botian Shi, Zhongying Tu, Conghui He

  46. NYU CTF Bench — Cotool evaluation, February 2026 ↗

    Cotool · 2026

  47. Terminal-Bench 2.0 leaderboard ↗

    Terminal-Bench Team

  48. AIME — evaluation methodology and results ↗

    Vals AI · 2026

  49. CaseLaw — evaluation methodology and results ↗

    Vals AI · 2026

  50. CorpFin — evaluation methodology and results ↗

    Vals AI · 2026

  51. GPQA Diamond — evaluation methodology and results ↗

    Vals AI · 2026

  52. LiveCodeBench — evaluation methodology and results ↗

    Vals AI · 2026

  53. LegalBench — evaluation methodology and results ↗

    Vals AI · 2026

  54. MedCode — evaluation methodology and results ↗

    Vals AI · 2026

  55. MedQA — evaluation methodology and results ↗

    Vals AI · 2026

  56. MedScribe — evaluation methodology and results ↗

    Vals AI · 2026

  57. MGSM — evaluation methodology and results ↗

    Vals AI · 2026

  58. MMLU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  59. MMMU Pro — evaluation methodology and results ↗

    Vals AI · 2026

  60. MortgageTax — evaluation methodology and results ↗

    Vals AI · 2026

  61. Poker Agent — evaluation methodology and results ↗

    Vals AI · 2026

  62. ProofBench — evaluation methodology and results ↗

    Vals AI · 2026

  63. Vals RSI Index — evaluation methodology and results ↗

    Vals AI · 2026

  64. SAGE — evaluation methodology and results ↗

    Vals AI · 2026

  65. SWE-bench Verified — evaluation methodology and results ↗

    Vals AI · 2026

  66. TaxEval — evaluation methodology and results ↗

    Vals AI · 2026

  67. Vibe Code Bench — evaluation methodology and results ↗

    Vals AI · 2026

Names used in source tables

GPT 5.2 · GPT 5.2 (Mini-SWE-agent) · GPT 5.2 (OpenHands) · GPT 5.2 (xhigh) · GPT-5.2 · GPT-5.2 (Codex) xhigh · GPT-5.2 (Medium) · GPT-5.2 (Xhigh) · GPT-5.2 (high) · GPT-5.2 (low) · GPT-5.2 (medium) · GPT-5.2 (unknown thinking) · GPT-5.2 (xhigh) · GPT5.2 · gpt-5.2 · gpt-5.2 (xhigh)